Opening the Curtain on Playwright Gender: An Integrated Economic Analysis of Discrimination in American Theater

Human capital theory emphasizes the contributions of differences in ability, education, experience, and training to differences in labor market outcomes between the genders.3 Perhaps the

Opening the Curtain on Playwright Gender:

An Integrated Economic Analysis of Discrimination in American Theater

BY

Emily Glassberg Sands

Submitted to Princeton University Department of Economics

April 15, 2009

I dedicate this thesis to my dear grandma Sunny, whose deep appreciation of education has long inspired my own,

whose endless generosity has opened countless doors throughout my life journey,

and whose unwavering love and friendship will always mean the world to me

ACKNOWLEDGEMENTS

I have sincere gratitude to people from multiple communities for contributing to this thesis; and I have eternal gratitude to this thesis for connecting me more deeply with all of you A community of Princeton professors, a community of New York playwrights, and a community of friends and family were particularly instrumental throughout the journey

Within the community of Princeton professors I am most grateful to my two advisers, Professor Cecilia Rouse and Professor Christina Paxson

To Professor Rouse, who advised me for the first four months of the project: Thank you for laughing with me when we threw out my umpteenth topic in favor of trying yet another one on for size Once settled on examining gender discrimination in the theater industry, thank you for working with me to develop a range of methodologies, teaching creativity by example By early October, I couldn‘t help but notice the extra little jump my step acquired on the way to see you each week Thank you for sharing in

Many thanks, too, to Professor Henry Farber, Professor Alan Krueger, and Professor Orley Ashenfelter To Professor Farber: Thank you for the econometric insights relayed during the grand tour of your brain; they were endlessly mind-expanding and wonderfully practical To Professor Krueger: Thank you for our general discussions on methodologies and for your contagious enthusiasm about the theater world; I look forward to seeing your and Julia‘s play To Professor Ashenfelter: Thank you for your thoughtful comments on the study of Broadway profitability; your creative ideas on controlling for production costs proved particularly useful In addition to the gratitude I feel for many of the members of Princeton University‘s Industrial Relations Section individually, I also have gratitude to the Section as a whole for generously funding this thesis

Within the community of New York playwrights I am most grateful to Julia Jordan and Sheri Wilner To Julia: Thank you for reaching out to Steve Levitt and, therefore, for indirectly reaching out to me; I am very grateful for such a wonderful topic, which opened a world full of delightful new friends Thank you for taking me under your wing and welcoming me unreservedly into your community To Sheri: Thank you for your endless fervor, even as we trudged through the nitty-gritty Thank you for sending along multiple articles a day to initiate me into the playwriting world, for answering

every theater question a girl could possibly have – usually within a matter of minutes, and for making me giggle – at Bake-Off and beyond To Julia and Sheri both: My thesis memories get no better than late-night Bar Centrale

I am grateful to several other New York playwrights as well Thank you to Lynn Nottage, Tanya Barfield, Julia Jordan, and Deb Laufer for each generously donating one script for this thesis‘s audit study Within the Dramatist Guild of America, special thanks

to Executive Director Gary Garrison for kindly providing the gender-coded membership list and, more generally, for his overarching support of this project; in addition, thanks to Ralph Sevush, the Dramatist Guild‘s Executive Director of Business Affairs, for his insights into the details of playwriting contracts

I also have huge gratitude to my friends and family for their unwavering support throughout the process Many thanks to Jonathan Mayer (‘09) for kindly putting his computer programming skills to work scraping data for use in this project, and to Kelley Taylor for wonderful proofreading in the final days To Holger Staude (‘09): Thank you for your insightful suggestions and excellent editing throughout this process; more generally, my Princeton experience has been all the more intellectual – and all the more joyful – in light of our ongoing discussions To Peter Capkovic (‘09), who was right there with me day in and day out: I feel very fortunate to be your person Finally, endless and loving thanks to my parents, Bob and Sally Sands, who have engaged with me extensively on this topic and countless others Dad and Mom: You two will forever be

my rocks

TABLE OF CONTENTS

Acknowledgements I

Chapter 1 Introduction: Hypotheses for the Paucity of Female-Written Plays 1

Chapter 2 Theory: The Economics of Labor Market Differences in Playwriting 6

2.1 Human Capital Theory 6

2.2 Discrimination Theory 9

2.2.1 Statistical Discrimination 11

2.2.2 Taste-Based Discrimination 13

2.3 A Model of Taste-based Discrimination in Theater 16

2.3.1 Discrimination by Audience Members 18

2.3.2 Discrimination by Cast and Crew 19

2.3.3 Discrimination by Artistic Directors 21

Chapter 3 Empirics: Methods of Testing for Discrimination in Theater 23

3.1 Separating the Explained from the Unexplained 23

3.2 An Audit Study Adapted 25

3.3 Chicago Price Theory on Broadway 27

Chapter 4 Separating the Explained from the Unexplained 29

4.1 Data 32

4.1.1 Worldwide Sample 34

4.1.2 American Sample 37

4.2 Methodology and Results 40

4.2.1 Worldwide Sample 41

4.2.2 American Sample 48

4.3 Conclusion 52

Chapter 5 An Audit Study Adapted 54

5.1 Experimental Design 55

5.1.1 Selecting Scripts 55

5.1.2 Identifying Recipients 55

5.1.3 Creating Pen-Names 56

5.1.4 Randomizing Survey Versions 58

5.1.5 Obtaining a High Response Rate 60

5.1.6 Targeting Gender Discrimination 61

5.2 Data 65

5.2.1 Respondent Characteristics 65

5.2.2 Outcome Variables 67

5.2.3 Revealed Importance of Variables 70

5.3 Methodology and Results 73

5.3.1 General Analysis 73

5.3.2 Extension: Comparison of Results by Respondent Gender 76

5.3.3 Extension: Comparison of Results by Protagonist Gender 80

5.4 Conclusion 85

Chapter 6 Chicago Price Theory on Broadway 87

6.1 Theory: Testing for Discrimination via Profits 87

6.2 Empirics: Testing for Discrimination via Profits 90

6.2.1 Methodological Issues 90

6.2.2 Data 92

6.2.3 Methodology and Results 94

6.3 Conclusion 102

Chapter 7 Conclusions and Recommendations for Further Analysis 104

Appendices 108

Appendix A Hard-Copy of Online Audit Study 108

Email to Survey Recipients 108

Online Survey 109

Appendix B Works Cited 162

Honor Pledge 167

CHAPTER 1

INTRODUCTION: HYPOTHESES FOR THE PAUCITY OF

FEMALE-WRITTEN PLAYS The vast majority of theater productions are written by men In the 2008/2009 New York Broadway season as it has been put forth, the percentage of plays written by women will amount to only 12.6% according to theaters‘ announcements This same figure one century ago, for the 1908/1909 New York season, was 12.8% according to the Internet Broadway Database Of productions in non-profit subscription houses with more than 99 seats in New York this year, moreover, only 17.8% were female-written.1Through discussions with playwrights, artistic directors, and producers in New York, I have collected a range of possible explanations for the consistently low number of female-written scripts chosen for production. 2

In labor economics terms, these explanations can be separated into human capital explanations and discrimination explanations Human capital theory emphasizes the contributions of differences in ability, education, experience, and training to differences

in labor market outcomes between the genders.3 Perhaps there are inherent differences in the abilities of men and women to write plays; some claim that female works lack the

―artistic exceptionalism‖ found in many male works Or it could be that men choose to

1 Data on the playwrights of productions in non-profit subscription houses in New York with more than 99 seats over the past year were compiled by Julia Jordan

2 Throughout this thesis, I will make the following distinction, widely accepted in the theater community:

 ―Script‖ refers to a written theatrical work

 ―Play‖ refers to a written theatrical work that has reached production

3 In this thesis, a theater ―employs‖ a playwright if it contracts with that playwright to produce one of his / her scripts While playwrights are, in practice, self-employed, I nonetheless use the term ―employment‖ to refer to this contracting process in order to facilitate smooth comparisons between this thesis and the preceding literature on employment discrimination

invest more in playwriting education and training than do women Although men and women appear to graduate from the top playwriting institutions in approximately equal proportions, this sample size is small and excludes the majority of American playwrights.4 Alternatively, women may be more likely to put playwriting on pause in order to, for example, bear and raise children; this may partially explain why many artistic directors note that the quantity of female-written submissions is lower than the quantity of male-written submissions

Human capital differences aside, discrimination may partially account for the limited number of female-written plays According to Heckman (1998), ―Discrimination

is a causal effect defined by a hypothetical ceteris paribus conceptual experiment –

varying [gender] but keeping all else constant.‖ Discrimination in playwriting would exist

if a female playwright is treated differently than an otherwise identical male playwright

The literature distinguishes between statistical discrimination and taste-based discrimination Statistical discrimination occurs when, in the presence of imperfect information about the actual productivity of workers, employers are forced to rely on visible features such as gender in judging each worker‘s productivity.Since there is near perfect information about the quality of the script as written, statistical discrimination hypotheses in playwriting focus on imperfect information about key characteristics of the playwright that may influence the success of the script in production; such characteristics include the playwright‘s ability to perform re-writes and to work effectively with the

artistic director

4 According to both the Yale School of Drama and the Juilliard School‘s Drama Division, the numbers of men and women enrolled in the programs have been approximately equal over the past decade

Taste-based discrimination, in contrast, occurs when customers, workers, or employers have a preference for interacting with one group of workers over another, despite equal productivity between the two groups of workers (Becker, 1971) In theater, customer discrimination would entail audience members choosing to buy tickets or pay higher ticket prices for a male-written play of the exact same type and quality as a female-written play purely because the former was written by a man; this would, in turn, drive down the revenues garnered by female-written plays Worker discrimination, meanwhile, would entail a preference by cast and crew for working with male playwrights; by demanding higher wages as compensation for working on a female-written play, discriminatory workers would drive up the variable costs of producing the female-written work Finally, employer discrimination would entail artistic directors preferring to work with male playwrights; given employer discrimination, artistic directors would select male-written scripts for production over female-written scripts, even amid equal profit potential for the two works in production

In this thesis, I first further develop and contextualize the human capital and discrimination hypotheses for the small number of female-written plays To that end, in the next chapter I explore the theoretical literature on human capital theory and discrimination theory before developing a theoretical framework with which to examine potential taste-based discrimination in theater This framework incorporates the possibility of all three sources of prejudice: discrimination by (1) audience members, (2) cast and crew, and (3) artistic directors

Building on this theoretical discussion, I present the three empirical methods employed throughout the remainder of this thesis in Chapter 3 On aggregate, the three

methods serve to disentangle the possible human capital and discrimination explanations for the small number of female-written plays Each method is grounded in the empirical literature on discrimination; each is also adapted in specific ways for applicability in the script-selection process

In Chapter 4, I use data on over 20,000 playwrights and 80,000 scripts, both produced and not yet produced, to separate the gap in the number of female-written and male-written plays into an explained gap and an unexplained gap The explained gap arises from differences in script characteristics, such as the number of parts, or roles, and the percentage of those parts that are female, as well as from differences in playwright characteristics, such as whether or not the playwright is represented by a literary agent The unexplained gap, in turn, is partially attributable to the gender of the playwright This method provides valuable insights into which script and playwright characteristics are most highly correlated with a script reaching production However, since some key characteristics of both the scripts and the playwrights are not observable, the results are likely subject to omitted variable bias

The second method, modeled after the classic audit study and presented in Chapter 5, overcomes any omitted variable bias by controlling fully for all characteristics

of the script, thereby isolating playwright gender I conducted an experiment that entailed sending four different scripts to artistic directors around the country, varying only the gender of the pen-name on each; I then collected information on artistic directors‘ assessments of the scripts I use these data to measure both statistical discrimination and taste-based discrimination in the script-selection process By asking artistic directors to rate the scripts along a variety of metrics, including play quality, economic prospects,

audience appeal, ease of casting, and fit with theater, results of this study illuminate not only any gender differences in whether or not scripts are selected for production, but also the motivations behind any gender differences

Since the experimental approach of the second method may give rise to hypothetical bias, I also examine potential discrimination in the script-selection process

by comparing the observed profits of male-written and female-written plays on Broadway over the past decade This method, presented in Chapter 6, is loosely analogous to comparing the marginal revenue products (MRP) of workers to their wages; here, a play‘s profits are substituted for the worker‘s MRP and whether or not a script reaches production is substituted for the worker‘s wage By examining profitability, as opposed to other potential measures of play quality, this approach precisely targets employer discrimination

in ability, education, experience, and training to differences in labor market outcomes between men and women Discrimination theory, in contrast, considers how statistical discrimination or pure prejudice may explain any differences between the genders These two theories are not mutually exclusive; much of discrimination theory examines whether prejudice can explain the residual difference in labor market outcomes of men and women once human capital characteristics are held constant

In what follows, I review aspects of human capital theory and discrimination theory relevant to the playwriting industry A discussion of the literature on these two theories, and a range of sub-theories within them, will contextualize most of the feasible explanations for the small number of female-written plays Grounded in the theoretical literature, I conclude this chapter by presenting a theoretical model of taste-based discrimination in theater

2.1 HUMAN CAPITAL THEORY

The small number of female-written plays is likely explained in part by human capital differences between the genders These human capital differences may arise from differences in endowments and/or investments In the latter case, either female

playwrights may choose to invest less in playwriting human capital, and therefore be less successful in their playwriting careers, or women may choose not to invest in playwriting human capital at all, thereby driving down the proportion of all playwrights who are female Any differences in investment choices between the genders may be driven by differences in ability, in costs and benefits of investments, or in preferences

A difference in human capital endowments between the genders such that men are more naturally adept than women at playwriting could account for the small number of female-written works in production The link between differences in human capital endowments and differences in labor market outcomes was perhaps first approached by Moore (1911) in his examination of the relationship between ability and earnings Upon testing the hypothesis that wage differences are attributable to ability differences, Moore concludes that ―[t]he natural difference in ability between individual laborers does find its expression in the apportionment of earnings among laborers in the present industrial state‖ (Moore, 1911) Since Moore, numerous economists have sought to model the

precise relationship between ability and wages (see, for example, Staehle, 1943); despite variations in their models, all agree that a relationship between ability and labor market outcomes exists

With or without differences in endowments, investment differences between the genders may explain part of the small number of female-written works in production In general, human capital investments include education, work experience, and job training

In playwriting, investments also include attendance at festivals, conferences, and other forums for discussing and developing scripts Even if men and women have identical innate playwriting abilities, they may still differ in the amount of human capital they

accumulate; gender differences in human capital investment would then be a matter of choice based on differential costs and benefits of investment, or on preferences, each of which is discussed below (Mincer, 1958).5

Female playwrights may choose to invest less than male playwrights in light of an expectation of spending less time playwriting over the course of their lives Altonji and Blank (1999) note that investment in market skills will be lower among those who plan to spend less time in the labor market Since many women expect to spend years bearing and raising children, women may choose to invest in less human capital than their male counterparts and, therefore, achieve poorer labor market outcomes

In light of differential benefits of investments, women may also be less likely to invest in playwriting human capital at all, instead selecting other professions As Polachek (1975) notes, since women‘s employment may be sporadic due to domestic responsibilities, women‘s job skills may depreciate more than men‘s; therefore, women

may choose jobs that have low penalties for sporadic employment This would, Polachek concludes, result in segregation even in the absence of any discrimination by employers

To the extent that success in playwriting demands a relatively consistent reputation and maintained connections with artistic directors, some women may believe that sporadic work is not conducive to becoming a successful playwright. 6

Women may also be less likely than men to invest in playwriting human capital due to a difference in preferences between the genders Noting that differences in the

preferences of men and women may lead to gender differences in wages, employment, and occupations, Altonji and Blank (1999) cite evidence that women are more likely than men to prefer professions with lower levels of personal risk If so, women may choose not to enter the playwriting profession as success in this occupation is highly variable and

difficult to predict ex ante

2.2 DISCRIMINATION THEOR Y

Human capital differences aside, discrimination may explain the small number of female-written plays Aigner and Cain (1976) divide discrimination into pre-market discrimination and labor market discrimination Pre-market discrimination refers to unequal opportunity to develop talents and abilities prior to employment; this includes inferior access to education and training In the case of playwriting, few have hypothesized that women face pre-market discrimination; as mentioned, men and women appear to graduate from at least the top playwriting institutions in approximately equal numbers Most discrimination-based explanations for the small number of female-written plays focus on labor market discrimination, which occurs upon entering the work force (Aigner and Cain, 1976) Labor market discrimination can be separated into wage, employment, and occupational discrimination

Wage discrimination would occur in playwriting if female playwrights were paid less than male playwrights for otherwise identical scripts.7 Wage discrimination in playwriting has, however, received little attention, likely due to the standardization of

7

Just as the term ―employment‖ is used throughout this thesis to refer to the contracting between theaters and self-employed playwrights, the term ―wage‖ is used to refer to the fee paid by theaters to playwrights in exchange for the rights to produce their scripts The common terms of ―employment‖ and ―wages‖ are substituted in for more precise theater terminology to facilitate parallels between this thesis and the preceding literature on discrimination

playwriting contracts endorsed by the Dramatist Guild (DG).8 All Guild members are strongly urged to use DG contracts, which specify, among other things, how much the writer is to be paid.9 According to Ralph Sevush, the DG‘s Executive Director of Business Affairs, who is responsible for setting up the contracts and their fee payment schemes, the payment levels vary only with where the show is being produced and with how many audience members can be seated at the producing theater These fixed fee schedules leave little room for wage discrimination.10

Employment and occupational discrimination are the focus of most discussions on discrimination in playwriting Amid employment discrimination, female playwrights would be less likely to attain production, holding the quality of scripts constant between the genders Amid occupational discrimination, moreover, women would be restricted from entering the playwriting occupation and/or would be crowded into other occupations despite equal playwriting capabilities between the genders Regardless of its type, labor market discrimination may take the form of statistical discrimination or taste-based discrimination In the latter case, a taste for male-written works could come from any combination of audience members, cast and crew, and artistic directors

8

According to its website, the DG is ―a community of playwrights, composers and lyricists dedicated to protecting, informing, and promoting the interests of dramatists everywhere.‖ The DG includes, among others, nearly every American playwright who has produced on a First-Class/Broadway, Off-Broadway, or mainstage of a regional theater (LORT) contract

9

Not all playwrights are compensated for their work exactly in line with the contracts At times, a playwright‘s agent (if he/she has one) negotiates a fee Other times, playwrights are very eager to get their work produced and will accept a lower fee – or even no money at all – for the opportunity to have a script reach production However, the DG strongly discourages its members from accepting substandard fees by removing membership privileges from any playwright who accepts any contract not approved by the Guild Therefore, despite some non-compliance, the contracts remain largely effective in standardizing fees

10 Although contracts hold playwright fees fixed for theaters of a given size in a given location, if written works are relegated to smaller theaters than otherwise-equivalent male-written works, indirect wage

female-2.2.1 STATISTICAL DISCRIMINATION

The theory of statistical discrimination provides an explanation of why rational employers might discriminate Phelps (1972) and Arrow (1973) assume that firms have limited information about the skills of applicants, and that easily observable characteristics like gender are correlated with performance, even after controlling for all the additional information firms have on potential workers After experiencing a correlation between the easily observable characteristics and productivity, employers use the easily observable characteristic as a proxy for the unobservable characteristics that would result in productivity differences (Arrow, 1973)

In playwriting, there is perhaps less imperfect information than in other professions; after all, the script itself is immediately, directly, and fully observable to artistic directors through the submission process According to artistic directors, moreover, the best predictor of the success of a play in production is ―what is on the page.‖ However, given the large number of script submissions artistic directors receive daily, they often do not have time to read each script in its entirety; this implies that imperfect information and, thus, statistical discrimination by artistic directors may arise even with regards to the written script

In addition, potential for imperfect information arises with regards to certain characteristics of the playwright According to artistic directors, the ―human component,‖ although less influential than what is on the page, can influence the success of a play in production In the script-selection process, artistic directors often consider the following: Will the playwright be capable of re-writes? Will he/she be easy to work with? Will he/she have a successful future career, thereby reflecting well on this theater in the years

to come? To the extent that such characteristics vary – or are perceived to vary – between the genders, or to the extent that such characteristics are harder to predict accurately in female playwrights than in male playwrights, statistical discrimination may arise

The literature on statistical discrimination identifies two routes through which it may operate In the first route, beliefs about the average skill of female playwrights as compared to the average skill of male playwrights may influence hiring decisions If prior beliefs about the productivity of female playwrights as compared to male playwrights are used in the decision making process, stereotypes may become self confirming and, therefore, self-perpetuating Coate and Loury (1993) demonstrate that prior beliefs can lead to equilibria in which groups with the same ability end up with different levels of skill In the case of playwriting, if theaters think that a smaller proportion of female playwrights are qualified, this will affect the human capital investment decisions of women and may thus confirm theaters‘ prior beliefs

In the second route, the actual productivity of female playwrights could be more difficult for theaters to predict than the actual productivity of male playwrights The resulting differences in ―riskiness‖ between female and male playwrights may then lead risk-averse theaters to discriminate statistically against female playwrights (see, for example, Aigner and Cain, 1977; Lundberg, 1991) As before, this type of statistical discrimination could affect the investment decisions of female playwrights and result in differences in productivity levels between the genders even in the face of previously equal average abilities

2.2.2 TASTE-BASED DISCRIMINATION

The theory of taste-based discrimination developed by Becker (1971) explains another potential cause of differences in labor market outcomes between male and female playwrights Although it is challenging to distinguish empirically from statistical

discrimination, taste-based discrimination operates quite differently As Becker defines it,

―if an individual has a 'taste for discrimination,' he must act as if he were willing to pay

something, either directly or in the form of a reduced income, to be associated with some persons instead of others When actual discrimination occurs, he must, in fact, either pay

or forfeit income for this privilege‖ (Becker, 1971) Becker discusses three sources of

prejudice: customers, workers, and employers

2.2.2.1 CUSTOMER DISCRIMINATION

Customer discrimination occurs when customers prefer to be served by one group

of workers as opposed to by another In Becker‘s model, firms that hire more of the disliked group cannot charge their customers as much and, therefore, pay lower wages Assume customers have tastes for discrimination against members of a particular

group If p is the money price of a unit of output produced by this group, a discriminatory customer would act as if p(1+d) were the net price, where d is the discrimination

coefficient of this customer In the absence of discrimination, two groups that are perfect substitutes in production would receive the same competitive equilibrium wage rate; customer discrimination against a particular group, however, reduces members of that group‘s wages relative to the wages of others (Becker, 1971)

Explanations for the small number of female-written plays rarely cite customer discrimination Customer discrimination would occur in theater only if audience

members chose to buy a ticket or pay a higher ticket price for a male-written play of the exact same type and quality as a female-written play purely on the basis of playwright gender With few exceptions, however, audience members are largely oblivious to a play‘s playwright – and thus to his or her gender As they used to say on Broadway,

―Only Mamet sells tickets,‖ implying that no other playwright‘s name drove sales.11

Playwright identity and, thus, playwright gender are seemingly unlikely to have a large impact on the utility audience members derive from seeing a given show Customer discrimination against female playwrights is therefore rarely discussed

2.2.2.2 WORKER DISCRIMINATION

Worker discrimination occurs when fellow workers prefer to work with workers

of one group over those of another A discriminatory worker‘s utility function is a function not only of goods and leisure, but also of the level of employment of workers in the disliked group Discriminatory workers therefore require higher wages to compensate for higher levels of employment of the disliked group within the firm; the discrimination

coefficient d, which measures the magnitude of the worker‘s taste for discrimination, converts a money wage rate w into a unit net wage rate w(1-d)

In theater, worker discrimination may arise if cast and crew members prefer to work on male-written productions In a recent email one artistic director writes, ―Star actors have A LOT of say I don't think it is going out on a limb to suggest that they want

to associate with the dominant force of theater to better their careers.‖ If this dominant

11 This year, even Mamet didn‘t sell tickets He had two plays on Broadway One flopped with bad reviews The other soared with good reviews until its TV star left; then it, too, plummeted Perhaps the new adage should be, ―Only Shakespeare sells tickets – and that‘s with a star and good reviews, and not so much without.‖

force is male, star actors may demand higher pay as compensation for working on female-written works Similarly, if set designers, costume designers, and other members

of the crew are more eager to work on male-written productions, they may demand higher wages for working on female-written productions

Theoretically, an income-maximizing employer faced with worker discrimination would never hire a mixed work force, as he/she would have to pay discriminatory workers working with workers belonging to the disliked group a higher wage rate than discriminatory workers working solely with other workers of the same group Rather, amid equal marginal revenue products (MRP) between the two groups of workers, the employer hires only the discriminatory group if their wage is less than that of the disliked group and only the disliked group if their wage is less than that of the discriminatory group Worker discrimination can therefore lead to segregation (Becker, 1971)

In the theater industry, partial segregation does occur The website of the Fund for Women Artists lists over thirty exclusively female theaters One of the most prominent is New George‘s in New York City As a self-described ―play and artist development

organization, providing essential resources to a community of venturesome artists (who are women),‖ New George‘s is about ―sparking opportunities and relationships that push adventurous artists (who are women) and their work forward into the world.‖ Although New George‘s hires the occasional male actor when a script demands it, the vast majority

of New George‘s workers are female Numerous other theaters mirror New George‘s

model of hiring predominately female workers and cultivating only female playwrights While I am not aware of any exclusively male theaters, worker discrimination could be one driver behind this partial segregation in the theater industry

2.2.2.3 EMPLOYER DISCRIMINATION

Employer discrimination occurs when the employer prefers to work with workers

of one group over workers of another, despite equal productivities between the two groups The employer's utility function, then, is a function not only of profits, but also of the percentage of workers who are in the disliked group In the market place, ―objective‖ employer behavior is based on considerations of productivity alone An employer discriminates by refusing to hire someone with a marginal value product greater than marginal cost; his/her behavior then lacks ―objectivity.‖ In Becker's analysis, when faced

with the money wage rate w a discriminatory employer acts as if w(1+d) were the net wage rate, with d being a discrimination coefficient measuring the intensity of the

employer‘s taste for discrimination

Employer discrimination in theater would entail discrimination by the artistic directors deciding which scripts to select for production From my discussions with artistic directors, I gather that many of them factor how well they will be able to relate to the playwright into the decision of whether to produce any given work An effective working relationship between the artistic director and the playwright are especially important in light of the re-write process, which necessitates close interaction between the two If artistic directors believe they can better relate to a male playwright than to an equally qualified female playwright, employer discrimination may arise

2.3 A MODEL OF TASTE-BASED DISCRIMINATION IN THEATER

Building on the previously discussed theoretical literature, I conclude this chapter

by presenting a model detailing the effects of each of the three potential sources of based discrimination on the selection of scripts for production

taste-At the most basic level, the total profits accrued by a theater for producing a script equal the total revenues minus the total costs Total revenues are equal to the price of a ticket times the number of seats sold per week times the number of weeks in production, where the price of a ticket is an increasing function of the quality of the play and a decreasing function of number of tickets.12 Total costs can be separated into variable costs and fixed costs Variable costs equal the number of weeks the play remains in production times the weekly costs of variable inputs such as labor and capital Fixed costs are equal to the fee paid to the playwright for the script plus the sum of all additional fixed costs such as set construction That is:

𝐸𝑞𝑢𝑎𝑡𝑖𝑜𝑛 2.1:

𝜋 = 𝑝 𝑞, 𝑛𝑠 𝑛𝑠 − 𝑐 𝑤, 𝑟, 𝑋 𝑛 − (𝑓 + 𝑍)

where 𝜋 indicates profits Within revenues, p is the price of each ticket, q is quality of the

production, n is the number of weeks in production, and s is the number of tickets sold each week Within costs, variable costs are a function of the wage rate, w, the cost of capital, r, and all other variable costs, X, multiplied by the number of weeks in production, n; of fixed costs, f represents the fee paid to the playwright and Z represents

the sum of all additional fixed costs

Let artistic directors be presented with two scripts, one written by a man and the other written by a woman In deciding which script to produce, profit-maximizing artistic directors compare pure profits If 𝜋𝑀− 𝜋𝐹 > 0, they select the male-written play for production; if 𝜋𝑀− 𝜋𝐹 < 0, they select the female-written play for production

Playwright gender aside, assume that the two scripts are identical Therefore, they are of precisely the same quality (𝑞𝑀 = 𝑞𝐹 = 𝑞) For simplicity, assume also that,

12 i.e the demand function slopes downward

because they are otherwise identical, the scripts would show in the same size theater for the same number of weeks (𝑛𝑀𝑠𝑀 = 𝑛𝐹𝑠𝐹 = 𝑛𝑠) Grounded in these assumptions, I

examine the implications of each of the three sources of taste-based discrimination discussed by Becker (1971) and adapted to the theater industry in this chapter

2.3.1 DISCRIMINATION BY AUDIENCE MEMBERS

First, assume cast and crew are indifferent between working on the male-written script or on the female-written script (𝑤𝑀 = 𝑤𝐹) and that artistic directors derive utility

only from profits – not from the gender of the playwright; however, audience members prefer to see the male-written play Specifically, audience members act as if they are paying 𝑝′𝑀 = 𝑝𝑀 to see the male-written play and 𝑝′𝐹 = 𝑝𝐹(1 + 𝑑𝑐) to see the female-

written play, where 𝑑𝑐 is, as in Becker‘s model, the discrimination coefficient, with the

subscript c denoting that the discrimination arises from customer tastes

Since the scripts are otherwise identical, audience members are otherwise

indifferent between the two plays and therefore set 𝑝′𝑀 = 𝑝′𝐹 such that 𝑝𝐹 = 𝑝𝑀

Substituting 𝑓𝑀 = 𝑓𝐹 = 𝑓 into Equation 2.3 reveals that, given higher audience

discrimination and equal fees, the male-written script will be chosen for production over the female-written script as the profit gain to theaters from this selection is as follows:

𝐸𝑞𝑢𝑎𝑡𝑖𝑜𝑛 2.4:

𝜋𝑀− 𝜋𝐹 = 𝑑𝑐𝑝𝑀

1 + 𝑑 𝑛𝑠

2.3.2 DISCRIMINATION BY CAST AND CREW

Assume next that audience members have no preference between the two plays (𝑝𝑀 = 𝑝𝐹 ) and that, as before, theaters seek to maximize profits; however, because cast

and crew prefer to work with the male playwright, the wages demanded by theater workers as compensation for working on the female-written script are higher than the wages demanded by theater workers to work on the male-written scripts Specifically,

𝑤𝐹 = (1 + 𝑑𝑤)𝑤𝑀 where 𝑑𝑤is, as before, the discrimination coefficient; the subscript w

denotes that the discrimination arises from worker tastes The profit differential then becomes

As before, assume fee-equalizing contracts such as those successfully endorsed by the

DG Equation 2.6 then simplifies to

𝐸𝑞𝑢𝑎𝑡𝑖𝑜𝑛 2.7:

𝜋𝑀− 𝜋𝐹 = (𝑐𝐹 𝑤𝑀(1 + 𝑑𝑤), 𝑟, 𝑋 − 𝑐𝑀 𝑤𝑀, 𝑟, 𝑋 )𝑛

Amid discrimination by cast and crew, and the equalization of playwright fees for otherwise-equivalent works, theaters would accrue equal profits from the male-written work and the female-written work only if capital and labor were perfect substitutes in production, (i.e only if discriminatory crew members could be entirely replaced by capital of equal cost) Cast and crew are, however, instrumental to the existence of the theater industry.13 Because labor and capital are not perfect substitutes in theater production, theaters will accrue higher profits by selecting the male-written script for production

How much profit a theater will gain by selecting the male-written script for production amid discrimination by cast and crew depends not only on the discrimination

13 Theater professions such as that of lighting director may eventually be made obsolete with advances in technology Nonetheless, at least until robots become commercialized, other professions such that of the stagehand will remain crucial to theater productions Perhaps more important still, until those robots can

coefficient, 𝑑𝑤, but also on the extent of the scope for substitution between capital and

labor in that theater For example, if there is no scope for substitution, the cost function is additive.14 Then, Equation 2.7 simplifies to

𝐸𝑞𝑢𝑎𝑡𝑖𝑜𝑛 2.8:

𝜋𝑚 − 𝜋𝑓 = 𝑑𝑤𝑤𝑚𝐿𝑛

where L is the number of man-hours of labor need each week in the production In all likelihood there is some, but not full, scope for substitution between capital and labor in the theater industry Therefore, the true profit differential will lie somewhere above zero and below this upper bound

2.3.3 DISCRIMINATION BY ARTISTIC DIRECTORS

Finally, assume that neither audience members nor cast and crew discriminate, but that the artistic director derives utility not only from profits, but also from the gender

of the playwright Then, an artistic director acts as if his/her theater accrues profits 𝜋′ of

14 i.e 𝑐 𝑤, 𝑟, 𝑋 = 𝑤 ∗ 𝐿 + 𝑟 ∗ 𝐾 + 𝐶 𝑋 where L is the number of man-hours of labor and K is the units

of capital needed each week for the production

Because theaters that discriminate due to artistic directors‘ tastes gain no additional revenue for discriminating, they will eventually go out of business as long as there is a competitive market in which the artistic directors‘ of other theaters do not share this taste for discrimination Similarly, theaters that discriminate due to tastes of cast and crew will cease to exist in a competitive market if other theaters have cast and crew without tastes for discrimination Therefore, all labor market discrimination that persists

in a competitive industry theoretically results from customer tastes (Nardinelli and Simon, 1990) As I discuss in Chapter 6, however, the theater industry may not be perfectly competitive In addition to audience discrimination, then, discrimination by cast and crew and/or by artistic directors may persist

3.1 SEPARATING THE EXPLA INED FROM THE UNEXPLAINED

One strand of literature on discrimination focuses on distinguishing an explained gap in wages and/or employment from an unexplained gap The explained gap accounts for any differences in human capital endowments and/or investments; in a range of professions, differences in human capital account for a large proportion of differences in labor market outcomes (see, for example, O‘Neill and Polachek, 1993; Becker and

Lindsay, 1994; Light and Ureta, 1995; Olsen and Sexton, 1996; Blau and Kahn, 1997; Sicherman, 1996; Blau, 1998) The unexplained gap, in turn, is presented as evidence of discrimination (see, for example, Madden, 1985; Blau and Ferber, 1987) In Chapter 4, I adapt this approach to the study of the script selection process in theater

Parallels can be drawn between the script-selection process and the hiring process Artistic directors (employers) are presented with a range of script excerpts (resumes) from playwrights (job applicants) From these excerpts, they select some smaller number

of scripts for a full reading (an interview) Following the full reading, they may choose to

meet with the playwright in person (second-round interview) prior to deciding whether or not to produce the script (hire)

In other regards, however, the script-selection process is distinct from the classic hiring process Importantly, the product produced by the playwright is discrete; a theater signs a contract with the self-employed playwright, paying the playwright a fee in return for the rights to produce one of his/her scripts As discussed in Chapter 2, some human capital characteristics of the playwright, such as how capable he/she is of re-writes and how easy he/she is to work, may impact the success of the play in production According

to artistic directors, however, characteristics of the script itself have by far the greatest bearing

With data on over 20,000 playwrights and 80,000 scripts, some of which have reached production and others of which have not, I distinguish an explained gap in the likelihood that a given script reaches production from an unexplained gap attributable to playwright gender The explained gap accounts for differences in the characteristics of scripts, including the total number of parts and the percentage of those parts that are female, as well as for differences in the characteristics of the playwright, such as whether

or not he/she is represented by a literary agent Holding these script and playwright characteristics constant, the unexplained gap accounts for differences in the probability that a script reaches production arising from the gender of the playwright

This methodology is not, however, a perfect test of discrimination; any gap may either overestimate or underestimate the effect of discrimination Since it is impossible to control perfectly for all script and playwright characteristics, the results almost certainly suffer from omitted variable bias In addition, if an expectation of discrimination causes

women to invest less in playwriting human capital or to avoid the occupation all together, the resulting gap will understate the true effects of discrimination Less biased estimates

of the extent of gender discrimination in the script-selection process will come from alternative empirical methods

3.2 AN AUDIT STUDY ADAPTED

Audit studies, pioneered by researchers at the Urban Institute in Washington, D.C

in the early 1990s, provide an alternative approach to studying discrimination Audit studies are particularly useful for finding evidence of discrimination in hiring decisions Because discrimination involves differences in labor market outcomes for individuals who differ only along one dimension, audit studies put forth otherwise identical potential workers who vary only in the group to which they belong Whereas inferences regarding discrimination drawn from differences in employment rates controlling for other observed characteristics can, as discussed, suffer from bias if there are differences between two groups of workers that are unobserved by the econometrician, an audit study overcomes this problem by completely matching the characteristics of the workers in different groups

One example of an audit study is the distribution to employers of resumes that are identical in all regards except the group to which the individuals belong For example, resumes may vary only in the race or gender of the potential worker Researchers then look for any differences across groups in the probability that the employer invites the potential worker for an interview (see, for example, Cross et al., 1990; Turner et al., 1991; Kenney and Wissoker, 1994; Neumark, 1996) In Chapter 5, I adapt this methodology to study gender discrimination in the script-selection process

Recall that an analogy can be drawn between a script excerpt and a resume; adapting the classic audit study to examine the script-selection process, I distributed four previously unseen scripts to artistic directors around the country, varying only the gender

of the pen-name on each script I asked respondents to rate the script along a range of different metrics, including measures of perceived overall quality, economic prospects, audience appeal, ease of casting, characteristics of the playwright, and fit with their theaters Comparing the ratings garnered by scripts with female pen-names to those garnered by otherwise identical scripts with male pen-names provides direct insights into taste-based discrimination by artistic directors as well as into taste-based discrimination

by audience members and cast and crew as perceived by artistic directors This approach also permits a test for statistical discrimination by artistic directors

As mentioned, statistical discrimination is generally challenging to distinguish from taste-based discrimination In his discussion of racial discrimination, Arrow (1998) writes the following:

―If there are a number of observable variables, such as quantity of education, then the hypothesis

of statistical discrimination implies that an estimate of wages based on these observables will be

significantly improved by adding race as a predictor But this is the same conclusion as arrived at

by the hypothesis of market-based discrimination based on taste.‖

To distinguish statistical discrimination, Arrow continues, one must be able to observe each worker‘s marginal productivity The literature on testing for statistical discrimination centers around two methods: The first uses observational data (see, for example, Neumark, 1999; Altonji and Pierret, 2001); the second uses controlled experiments (see, for example, Fershtman and Gneezy, 2001; Dickenson and Oaxaca, 2006) This thesis‘s audit study is a controlled experiment

In theater, the submission process leaves little room for imperfect information with regards to the script unless the submitted script is not read in its entirety; the real potential for imperfect information, and thus for statistical discrimination, arises with regards to certain characteristics of the playwright important in the production process (see Chapter 2 for a more involved discussion) These characteristics include whether or not the playwright will be capable of re-writes, how easy he/she will be to work with, and how successful his/her future career will be To the extent that these characteristics vary –

or are perceived to vary – between the genders, or to the extent that these characteristics are harder to predict accurately in female playwrights than in male playwrights, statistical discrimination may arise With questions targeting these playwright characteristics in the audit study, I test for statistical discrimination by artistic directors

Although it is conducive to testing for both statistical discrimination and all three sources of taste-based discrimination in theater, the audit study presented in Chapter 5 may suffer from hypothetical bias To overcome the potential hypothetical bias, I employ

an observational approach to testing for gender discrimination in the script-selection process in Chapter 6

3.3 CHICAGO PRICE THEORY ON BROADWAY

Comparing the marginal revenue products of workers to their wages is an alternative method for detecting discrimination Whereas audit studies look at the effects

of discrimination on hiring, this approach examines the effects of discrimination on wages Derived from Becker‘s (1971) model of employer discrimination, evidence that one group of workers receives wages less than their marginal revenue product (MRP) while another group of workers receives wages equal to their MRP is evidence of

discrimination against the first group The method has been employed in a variety of professions, including professional sports (see, for example, Kahn and Sherer, 1988; Nardinelli and Simon, 1990; and Hellerstein et al., 1999) In Chapter 6, I adapt this methodology to the script-selection process

I substitute whether the script reaches production (employment) for wages because, as discussed in Chapter 2, the standardization of playwriting contracts endorsed

by the Dramatist Guild leaves little room for discrimination in payment schemes In addition, I substitute the quality of the play for the MRP of the worker since a playwright produces a discrete good I then compare the quality of female-written scripts selected for production on Broadway over the past decade to the quality of their male-written counterparts More specifically, defining quality as profits, I compare weekly revenues and run lengths of female-written and male-written productions on Broadway over the past decade, controlling both for the time of production and for the type of play, a partial control for production costs A test of the null hypothesis that female-written and male-written plays are equally profitably is a test of no employer discrimination

CHAPTER 4

SEPARATING THE EXPLAINED FROM THE UNEXPLAINED

In light of the small number of female-written works in production, some cite employment discrimination as a primary driver The small number of female-written productions is not, however, compelling evidence of employment discrimination, or even

of employment differences between the genders Perhaps women choose to enter the playwriting profession in lower numbers; the resulting occupational differences alone could explain the small number of female written works.15 Alternatively, or in addition, perhaps women who become playwrights choose to write fewer scripts; this could also result in a smaller number of female-written productions, even amid equal rates of production between female-written and male-written scripts Therefore, to set the stage for subsequent examination of discrimination in playwriting, this study examines occupational and employment differences between male and female playwrights using data from Doollee.com, an online database with records for over 20,000 playwrights and 80,000 scripts.16

15 This thesis does not examine occupational discrimination directly However, differences in the rates at which men and women enter and/or remain in the playwriting profession could be driven by a variety of factors, one of which is occupational discrimination If women perceive gender barriers in playwriting, they may choose not to enter the profession; in addition, amid employment discrimination discouraged female playwrights may exit the profession

16

This chapter uses data collected from Doollee.com Doolle.com is an online database of playwrights and their theater plays founded and maintained by Julian Oddy In response to an email request for a compilation of the data on Doollee, Mr Oddy responded, ―I am afraid I was a little bemused to receive your email asking me to send you a copy of a database with the fruits of years of research on it.‖ Nonetheless, the database is publically available and a computer program kindly created by Jonathan Mayer (‘09) effectively scraped the data from the website, creating an equivalent database for use in this thesis

Even if female playwrights do happen to be less likely to have their works reach production, this is not in and of itself evidence of employment discrimination; after all, male-written and female-written scripts likely differ in ways other than merely playwright gender In this chapter I test the extent to which any employment differences between the genders can be explained by gender differences in either or both of two important characteristics of scripts In addition, male and female playwrights may differ along more dimensions than gender, including representation by a literary agent In this chapter, I also test the extent to which employment differences between the genders can

be explained by gender differences in whether or not a playwright is represented by a literary agent

The first characteristic that may differ between female-written and male-written scripts, and which may lead to a difference in the rate at which female-written and male-written works reach production, is the proportion of parts that are female At a Town Hall Meeting on October 27, 2008 to discuss potential discrimination against female playwrights at New Dramatists in New York, one artistic director hypothesized that female playwrights are less likely to have their work produced because ―women tend to write about other women, who are harder for men to relate to Men have always had more difficulty relating to female characters than have women had difficulty relating to male characters; little boys will not watch stories about girls, but girls will watch stories about boys.‖ In this chapter, I distinguish between plays with majority female parts and plays

with majority male parts I then test whether female playwrights are more likely to write scripts with majority female parts and whether scripts with majority female parts are less likely to reach production

Second, at this same Town Hall Meeting, a few of the female playwrights noted that they feel compelled to write scripts with smaller casts in order to increase the likelihood of their scripts reaching production Intuitively, scripts with fewer parts are more likely to reach production because they are generally less expensive to produce and can be produced in any one of many smaller theaters In this chapter, I test whether women tend to write scripts with fewer parts and whether scripts with fewer parts are more likely to attain production

Finally, in outside discussions with other female playwrights, it was frequently mentioned that some literary agents seem to be more eager to represent male playwrights than to represent female playwrights Since many theaters only accept script submissions through literary agents, moreover, agents often serve as gatekeepers Even in the absence

of discrimination by artistic directors, discrimination by literary agents could lead to a smaller proportion of female-written scripts reaching production In this chapter, I test whether female playwrights are less likely to be represented by agents and whether a script whose playwright is represented by a literary agent is more likely to reach production

I examine both occupational and employment differences first at the play level and then at the playwright level At the play level, I test if women are less likely than men

to write scripts, if the scripts women do write are less likely to reach production, and to what extent any differences in the likelihood of reaching production can be explained by differences in whether most of the parts are female, in the total number of parts, and in whether or not the playwright is represented by a literary agent At the playwright level, I test if women are less likely than men to write at least one script, if women who do write

are less likely than their male counterparts to have at least one script produced, and whether any differences between male and female playwrights in the rates of first-production can be explained by the previously discussed script and playwright characteristics I perform all of these tests first on the full sample, which includes a wide range of plays written in the English language since 1956, and then on a sample restricted

to American playwrights registered with the Dramatist Guild

4.1 DATA

Data for this chapter were collected from Doollee.com, an online database of playwrights and theater plays.Founded in 2003 by Julian Oddy, Doollee is an ongoing project that seeks to list every script written or produced in English since 1956, the year

John Osborne‘s seminal script Look Back in Anger was first performed in the Royal

Court of London At the time of its launch, Doollee contained records for 3,093 playwrights and 8,146 scripts With daily self-reported submissions and additions made

by Julian Oddy himself throughout the years, the number of records on Doollee has ballooned since its 2003 founding On November 4, 2008, the date on which data for this paper were scraped from the website, Doollee had records for 20,447 playwrights and 81,330 scripts

Although Doollee is the most comprehensive database available, several cautions are in order First, although the database is quite extensive, it certainly does not contain records for every script written in English since 1956 Second, because Doollee contains

a collection of works chosen by Julian Oddy or self-reported by playwrights, those scripts that are included are likely not a random sample of all scripts written in English since

female-written scripts, my estimates of the proportion of all scripts written by each gender will be biased Perhaps more important, to the extent that male playwrights and female playwrights differ in the likelihood with which they submit records to the database, self-reporting will result in biased estimates

Third, records also vary in completeness and, presumably, in the frequency with which they are updated All records contain the name of the playwright followed by a list

of scripts written by that playwright; the list of scripts for each playwright is not, however, always fully comprehensive Many records also contain the name of the playwright‘s literary agent In addition, some records contain the number of male parts,

the number of female parts, and, if produced, the year of first production for each script

For the purpose of this study, I assume that playwrights without a listed literary agent have no literary agent It is possible, however, that some playwrights do not list their literary agent and/or that some playwrights obtained a literary agent after the initial record was created and that the record has not since been updated To the extent that male playwrights and female playwrights with literary agents differ in the rates at which they report having a literary agent or at which they update their profiles, my estimates of the rates at which male and female playwrights are represented by literary agents will be biased

Similarly, I assume that scripts without a year of first production have yet to be produced It is possible, however, that some of these scripts were produced after the initial submission of the record and that the record has not since been updated Again, if male playwrights differ from female playwrights in the frequency with which they update

the records on their works, my estimates of the rates at which male-written and written works are produced will be biased

female-Data limitations, then, pose several difficulties for my analyses While inspection

of the Doollee database provides some interesting insights into the attributes of the scripts that reach production as compared to the attributes of scripts that do not, the findings must be considered cognizant of the incompleteness of the data and of the likely sample selection problems

4.1.1 WORLDWIDE SAMPLE

Recall that the mission of Doollee is to include all plays written in English since

1956 Created by an Englishman, the database includes records of plays written by English-speaking playwrights from around the world Although I am most interested in studying the script-selection process in America, the worldwide characteristic provides valuable information – and a large sample size; I therefore include English-speaking playwrights from around the world in my first analysis

A key variable in this paper is the gender of the playwright From the Doollee records, I do not have explicit information on playwright gender I do, however, have data on each playwright‘s full name Using data on naming frequency from the 1990 US Census, I compute the probability that each of the high-frequency first names in the 1990

US Census is female as follows

𝐹𝑒𝑚𝑎𝑙𝑒𝐹𝑟𝑒𝑞𝑢𝑒𝑛𝑐𝑦𝑖 + 𝑀𝑎𝑙𝑒𝐹𝑟𝑒𝑞𝑢𝑒𝑛𝑐𝑦𝑖